Prasad Rajesh Posture

Batch: June 2022
Data Analytics with Python

Task : Peform data analysis on India's statewise covid-19 data using various visualization tools.

import libraries

import numpy as np
import pandas as pd
from pandas.plotting import scatter_matrix
import matplotlib.pyplot as plt
import seaborn as sns
sns.set_style('darkgrid')
import plotly.express as px
import warnings
warnings.filterwarnings("ignore")

Loading the data

Datasource : https://www.kaggle.com/datasets/prasadposture121/covid19-india-statewise-data

data= pd.read_csv('Covid-19_India.csv')
data.head()

	State/UTs	Total Cases	Active	Discharged	Deaths	Active Ratio	Discharge Ratio	Death Ratio	Population
0	Andaman and Nicobar	10039	0	9910	129	0.00	98.72	1.28	380581
1	Andhra Pradesh	2319869	62	2305076	14731	0.00	99.36	0.63	49577103
2	Arunachal Pradesh	64504	1	64207	296	0.00	99.54	0.46	1383727
3	Assam	724225	4	716235	7986	0.00	98.90	1.10	31205576
4	Bihar	830702	43	818403	12256	0.01	98.52	1.48	104099452

Renaming the data columns

data.rename(columns = {'Active Ratio':'Active Ratio (%)', 'Discharge Ratio': 'Discharge Ratio (%)', 'Death Ratio': 'Death Ratio (%)'}, inplace = True)

Exploratory Data Analysis

Various data types of the dataset

data.dtypes

State/UTs               object
Total Cases              int64
Active                   int64
Discharged               int64
Deaths                   int64
Active Ratio (%)       float64
Discharge Ratio (%)    float64
Death Ratio (%)        float64
Population               int64
dtype: object

Columns of the data

data.columns

Index(['State/UTs', 'Total Cases', 'Active', 'Discharged', 'Deaths',
       'Active Ratio (%)', 'Discharge Ratio (%)', 'Death Ratio (%)',
       'Population'],
      dtype='object')

Checking fot the null values

data.isnull().sum()

State/UTs              0
Total Cases            0
Active                 0
Discharged             0
Deaths                 0
Active Ratio (%)       0
Discharge Ratio (%)    0
Death Ratio (%)        0
Population             0
dtype: int64

Checking for the duplicate values

data.duplicated().sum()

0

Shape of the dataset (rows, columns)

data.shape

(36, 9)

Size of the dataset

data.size

324

Basic Information of the dataset

data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 36 entries, 0 to 35
Data columns (total 9 columns):
 #   Column               Non-Null Count  Dtype  
---  ------               --------------  -----  
 0   State/UTs            36 non-null     object 
 1   Total Cases          36 non-null     int64  
 2   Active               36 non-null     int64  
 3   Discharged           36 non-null     int64  
 4   Deaths               36 non-null     int64  
 5   Active Ratio (%)     36 non-null     float64
 6   Discharge Ratio (%)  36 non-null     float64
 7   Death Ratio (%)      36 non-null     float64
 8   Population           36 non-null     int64  
dtypes: float64(3), int64(5), object(1)
memory usage: 2.7+ KB

Stastical information of each column

data.describe()

	Total Cases	Active	Discharged	Deaths	Active Ratio (%)	Discharge Ratio (%)	Death Ratio (%)	Population
count	3.600000e+01	36.000000	3.600000e+01	36.000000	36.000000	36.000000	36.000000	3.600000e+01
mean	1.198233e+06	415.416667	1.183250e+06	14567.027778	0.026389	98.845000	1.128056	3.362689e+07
std	1.771219e+06	797.343424	1.745163e+06	26960.635812	0.033050	0.494221	0.491440	4.305758e+07
min	1.003900e+04	0.000000	9.910000e+03	4.000000	0.000000	97.650000	0.030000	6.447300e+04
25%	9.912050e+04	4.000000	9.802975e+04	1104.500000	0.000000	98.520000	0.870000	1.439840e+06
50%	5.892080e+05	69.000000	5.828095e+05	6505.500000	0.010000	98.860000	1.100000	2.106970e+07
75%	1.285924e+06	385.250000	1.276082e+06	14208.250000	0.040000	99.112500	1.397500	5.229275e+07
max	7.882476e+06	3799.000000	7.732792e+06	147856.000000	0.120000	99.970000	2.340000	1.998123e+08

See if there is any correlation between any data columns

Numeric Correlations

data.corr()

	Total Cases	Active	Discharged	Deaths	Active Ratio (%)	Discharge Ratio (%)	Death Ratio (%)	Population
Total Cases	1.000000	0.793212	0.999987	0.943813	0.164050	-0.135375	0.125179	0.533583
Active	0.793212	1.000000	0.794030	0.684131	0.535411	-0.121842	0.086232	0.277804
Discharged	0.999987	0.794030	1.000000	0.942144	0.164359	-0.133217	0.122980	0.533843
Deaths	0.943813	0.684131	0.942144	1.000000	0.122715	-0.266886	0.260741	0.490676
Active Ratio (%)	0.164050	0.535411	0.164359	0.122715	1.000000	-0.147545	0.079946	0.040004
Discharge Ratio (%)	-0.135375	-0.121842	-0.133217	-0.266886	-0.147545	1.000000	-0.997654	-0.058038
Death Ratio (%)	0.125179	0.086232	0.122980	0.260741	0.079946	-0.997654	1.000000	0.057243
Population	0.533583	0.277804	0.533843	0.490676	0.040004	-0.058038	0.057243	1.000000

Corrrelation using Heatmap

plt.figure(figsize=(6,6))
sns.heatmap(data.corr())
plt.title("Correlation Between the Datacolumns")

Text(0.5, 1.0, 'Correlation Between the Datacolumns')

Correlation using Scatter Matrix

scatter_matrix(data,figsize=[15,15],diagonal='kde')
plt.show()

Conclusion: It is evident from these plots and the numerical data of correlation that the total number cases and number of discharged people show high positive correlation therefore lot of people are getting recovered from this disease. This is also supported by the negative correlation between death ratio and discharge ratio. Although there is a significant amount of correlation between number deaths and total cases it is because of the exponetial spread of the virus.

Data Visualization

Comparing each attribute of the dataset such as total number of cases, number of active cases, deaths etc with different states / union territories.

#Getting the different attributes of the dataset
attributes=list(data.columns)
attributes.remove('State/UTs')
attributes

['Total Cases',
 'Active',
 'Discharged',
 'Deaths',
 'Active Ratio (%)',
 'Discharge Ratio (%)',
 'Death Ratio (%)',
 'Population']

for attribute in attributes:
    data=data.sort_values(attribute,ascending=False)
    plt.figure(figsize=(10,10))
    sns.barplot(x='State/UTs',y=attribute,palette='CMRmap',data=data)
    plt.title(attribute+" per State/UTs ",fontsize=15)
    plt.xticks(rotation=90)
    plt.show();

Conclusion: Maharashtra has the highest number of total cases and Andman and Nicobar are at the last place. Kerala has the highest number of active cases. Number of discharged people is large in Maharashtra but number of deaths are large as well that is because of the large and dense population in the different cities of Maharashtra. Haryana has the highest active case ratio and Punjab has the highest death ratio but the discharge ratio is almost same for each and every state / union territory. Finally the last plot gives population of each state / union territory.

One can make this plot more interesting by comparing the attributes of states / union territories on the map of India.

for attribute in attributes:
    fig=px.choropleth(data,
             geojson="https://gist.githubusercontent.com/jbrobst/56c13bbbf9d97d187fea01ca62ea5112/raw/e388c4cae20aa53cb5090210a42ebb9b765c0a36/india_states.geojson",
             featureidkey='properties.ST_NM',   
             locations='State/UTs',       
              color=attribute,
              color_continuous_scale='Inferno',
               title=attribute+' per State / UTs' ,  
               height=700
              )
    fig.update_geos(fitbounds="locations", visible=False)
    fig.show()
#                                                                                               Scroll Down From Here >>>>>>>

The above plots show same comparitive relationship between states/UTs and the attributes like the barplots, but with the map we can easily understand how the virus might have spread across the contry. The limitation of this plot is due smaller size union territories are not easily visible.

Getting the total number of cases,number of active cases , number deaths and discharges across the nation

data2=data.sum(axis=0)
data2[['Total Cases','Active','Discharged','Deaths']]

Total Cases    43136371
Active            14955
Discharged     42597003
Deaths           524413
dtype: object

The End